Model trust region technique in parallel Newton's method for training feedforward neural networks.
M. D. ZhaoX. WangPublished in: ISCAS (1993)
Keyphrases
- feedforward neural networks
- trust region
- probabilistic model
- prediction model
- newton method
- training algorithm
- cost function
- objective function
- linear model
- training samples
- log likelihood
- training process
- bp neural network
- optimization method
- back propagation
- simulated annealing
- sensitivity analysis
- training phase
- support vector machine
- em algorithm
- recurrent neural networks
- convergence rate
- optimization methods
- neural nets
- least squares
- dynamic programming